I had created a testcase for HSEARCH-638 [1] quite some time ago, and Davide was now so kind to take it over and implement the solution which is almost ready in a pull request.
In my opinion, HSEARCH-638 is a performance bug and should be fixed. When we last talked about it, Emmanuel was not considering it being bug, as it was exactly expressing it's original intention.. I believe we settled by saying it would be fine to change the behaviour, BUT he would not consider it a bug fix but an API change as it would affect possible users: indeed that's correct, as it changes what is exactly included in the index, but how I was seeing the index mapping I doubt anybody was relying on it. So now we have a fix, and the project lead considers it an API change. This was the original reason for me when I made the tests for it, to not provide a solution yet, as it would need to wait for 4.0 .. now we have the solution so I think we should rather merge it now than to wait as it might be considered an API change (although I'm not overly convinced of that). Problem is that it's a bit late in the 4.0 train. If you think it's acceptable to make another RC soon, then I'd love to include this and also propose a solution for HSEARCH-598 (MassIndexer freezes when pool size is too low), as I have a POC for it but it's not trivial so I won't send a pull request (nor spend more time on it) unless we'll make another RC. If we could, I'd include two other API changes: - HSEARCH-800 Should we move ORM related APIs to org.hibernate.search.orm? Today they are in org.hibernate.search - HSEARCH-795 Move some of SearchFactoryImplementor from spi to impl (and move SPI level contracts up) should these otherwise be rejected? Sanne 1 - Limit graph traversal by @ContainedIn to the minimum required path - https://hibernate.onjira.com/browse/HSEARCH-638 _______________________________________________ hibernate-dev mailing list hibernate-dev@lists.jboss.org https://lists.jboss.org/mailman/listinfo/hibernate-dev
